We have investigated the involvement of the b-catenin gene in hepatocarcinogenesis, we used several transgenic mouse models of hepatic tumors induced by over-expression of c-myc in the liver either alone or in combination with TGF-a or TGF-b1. Activation of b-catenin, as judged by the presence of mutations and/or nuclear translocation of the protein, was most frequent in liver tumors from c-myc (4/17, 23.5%) and c-myc/TGF-b1 (6/18, 33.3%) transgenic mice. However, it was very rare in faster growing and histologically more aggressive hepatocellular carcinomas developed in c-myc/TGF-a mice (1/20, 5%). Administration of diethylnitrosamine, phenobarbital or 2-amino-3,8-diethylimidazo[4,5-f]quinoxaline (MeIQx) to the c-myc mouse did not significantly affect the occurrence of b-catenin mutations. Notably, nuclear accumulation of b-catenin was observed only in adenomas and highly-differentiated carcinomas with eosinophilic phenotype. Furthermore, preneoplastic lesions with eosinophilic phenotype frequently displayed focal nuclear positivity, co-localized with areas of high proliferation. In contrast, basophilic and clear-cell foci, as well as pseudo-glandular and poorly-differentiated hepatocellular carcinomas, exhibited a normal or reduced membranous immunoreactivity for b-catenin. These studies suggest that nuclear translocation of b-catenin and activation of Wingless/Wnt signaling may represent an early event in liver carcinogenesis, providing a growth advantage in a subset of hepatic tumors with a more differentiated phenotype. This is of considerable interest since mutations in b-catenin are associated with a decreased loss of heterozygosity in human HCC). We have therefore hypothesized that development of liver cancer may advance via two broad pathways. The first pathway would involve b-catenin activation, while the second one would generate genomic instability and produce a mutator phenotype. We are currently investigating correlation between b-catenin activation and genomic instability by analyzing the genomic profiles of mouse liver tumors by the random amplified polymorphic DNA (RAPD) analysis using 22 arbitrary primers. The results so far have revealed that genomic instability was absent in c-myc and was very low in c-myc/TGF-b1 (5.8 ? 1.3%) tumors which displayed b-catenin activation. In contrast, 10 out of 10 HCCs developed in c-myc/TGF-a mice exhibited a high rate of genomic instability (25 ? 3.0%). The latter was comparable with the rate of genomic instability found in c-myc/ lacZ HCCs after treatment with the potent hepatocarcinogen 2-amino-3,8-dimethyl-imidazo[4,5-f]quinoxaline (32.3 ? 2.4%). The frequency of genomic instability was inversely correlated with the degree of HCC differentiation. Furthermore, loss of heterozygosity of b-catenin and TGF-a RII genes was detected only in c-myc/TGF-a (60%) and c-myc/lacZ (80%) poorly-differentiated HCCs, presumably representing a marker of tumor progression in both transgenic models. Taken together, the present data are consistent with the existence of two pathways of hepatocarcinogenesis in the transgenic mouse models. The first pathway would imply the cooperation between c-myc and b-catenin oncogenes in the presence of a stable genome, while the second pathway is operative in a context of genomic instability.